[PATCH] fix NUMA interleaving for huge pages
commit3b98b087fc2daab67518d2baa8aef19a6ad82723
authorNishanth Aravamudan <nacc@us.ibm.com>
Fri, 1 Sep 2006 04:27:53 +0000 (31 21:27 -0700)
committerLinus Torvalds <torvalds@g5.osdl.org>
Fri, 1 Sep 2006 18:39:10 +0000 (1 11:39 -0700)
treea7defc8fa53b2023affc072cd20c4f4734e4395d
parent1678df37be8abbb381becdc40242ed915e775550
[PATCH] fix NUMA interleaving for huge pages

Since vma->vm_pgoff is in units of smallpages, VMAs for huge pages have the
lower HPAGE_SHIFT - PAGE_SHIFT bits always cleared, which results in badd
offsets to the interleave functions.  Take this difference from small pages
into account when calculating the offset.  This does add a 0-bit shift into
the small-page path (via alloc_page_vma()), but I think that is negligible.
 Also add a BUG_ON to prevent the offset from growing due to a negative
right-shift, which probably shouldn't be allowed anyways.

Tested on an 8-memory node ppc64 NUMA box and got the interleaving I
expected.

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Adam Litke <agl@us.ibm.com>
Cc: Andi Kleen <ak@muc.de>
Acked-by: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
mm/mempolicy.c